Logical Foundations for Data Interlinking with Keys and Link Keys

Tracking #: 1893-3106

This paper is currently under review
Manuel Atencia
Jérôme David
Jérôme Euzenat

Responsible editor: 
Bernardo Cuenca Grau

Submission type: 
Full Paper
Both keys and their generalisation, link keys, have been proposed as a means to perform data interlinking, i.e. finding identical resources in different RDF datasets. However, the usage of keys and link keys for data interlinking has not been formalised yet. This is necessary to ensure the correctness of data interlinking tools based on keys or link keys. Furthermore, such a formalisation allows to understand the differences between keys and link keys and to pin down the conditions under which keys and link keys are equivalent. In this paper, we first formalise how keys can be combined with ontology alignments for data interlinking. Then, we extend the definition of a link key by giving the formal semantics of six kinds of link keys: weak, plain and strong link keys, and their in- and eq-variants. Moreover, we establish the conditions under which link keys are equivalent to keys. Finally, we logically ground the usage of these link keys for data interlinking and show that data interlinking with keys and alignments can be reduced to data interlinking with link keys, but not the other way around.
Full PDF Version: 
Under Review